Professional Documents
Culture Documents
Textbook Winning Ways For Your Mathematical Plays Vol 2 Second Edition Berlekamp Ebook All Chapter PDF
Textbook Winning Ways For Your Mathematical Plays Vol 2 Second Edition Berlekamp Ebook All Chapter PDF
https://textbookfull.com/product/biota-grow-2c-gather-2c-cook-
loucas/
https://textbookfull.com/product/the-winning-way-2-0-learnings-
from-sport-for-managers-anita-bhogle/
https://textbookfull.com/product/joyful-living-101-ways-to-
transform-your-spirit-revitalize-your-life-first-edition-mercree/
https://textbookfull.com/product/manufacturing-technology-
vol-2-4th-edition-rao/
Assassination Classroom Vol 2 Yusei Matsui
https://textbookfull.com/product/assassination-classroom-
vol-2-yusei-matsui/
https://textbookfull.com/product/turbostrategy-21-powerful-ways-
to-transform-your-business-and-boost-your-profits-quickly-tracy/
https://textbookfull.com/product/shaman-king-comixology-edition-
vol-2-hiroyuki-takei/
Second chance for your money your life and our world
Kiyosaki
https://textbookfull.com/product/second-chance-for-your-money-
your-life-and-our-world-kiyosaki/
https://textbookfull.com/product/the-ultimate-retirement-guide-
for-50-winning-strategies-to-make-your-money-last-a-lifetime-1st-
edition-suze-orman/
Winning Ways for Your Mathematical Plays, Volume 2
Winning Ways
for Your Mathematical Plays
Ɔ
Volume 2, Second Edition
A K Peters
Natick, Massachusetts
Editorial, Sales, and Customer Service Office
A K Peters, Ltd.
63 South Avenue
Natick, MA 01760
All rights reserved. No part of the material protected by this copyright notice may be reproduced or
utilized in any form, electronic or mechanical, including photocopying, recording, or by any informa-
tion storage and retrieval system, without written permission from the copyright owner.
Berlekamp, Elwyn R.
Winning Ways for your mathematical plays / Elwyn Berlekamp, John H. Conway,
Richard Guy.--2nd ed.
p. cm.
Includes bibliographical references and index.
ISBN 1-56881-130-6 (v. 1) – ISBN 1-56881-142-X (v. 2) – ISBN 1-56881-143-8 (v. 3) –
ISBN 1-56881-144-6 (v. 4) (alk.paper)
1. Mathematical recreations. I. Conway, John Horton. II. Guy Richard K. III. Title.
Printed in Canada
07 06 05 04 03 10 9 8 7 6 5 4 3 2 1
To Martin Gardner
who has brought more mathematics to more millions than anyone else
Elwyn Berlekamp was born in Dover, Ohio, on September 6, 1940. He
has been Professor of Mathematics and of Electrical Engineering/Com-
puter Science at UC Berkeley since 1971. He has also been active in
several technology business ventures. In addition to writing many jour-
nal articles and several books, Berlekamp also has 12 patented inven-
tions, mostly dealing with algorithms for synchronization and error
correction.
He continues to climb mountains with his wife, Louise, and they have
been patrons of the Association of Canadian Mountain Guides’ Ball
and recipients of the A. O. Wheeler award for Service to the Alpine
Club of Canada.
Contents
ix
x Contents ♥
Falada . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312
Two More Falada Games . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320
Baked Alaska . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
A Felicitous Falada Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324
The Rules for Tallies on Infinite Tolls . . . . . . . . . . . . . . . . . . . . . . . . . . . 325
Time May Be Shorther than You Think! . . . . . . . . . . . . . . . . . . . . . . . . . . 326
Glossary 453
Index 455
Preface to the Second Edition
It’s high time that there was a second edition of Winning Ways.
Largely as a result of the first edition, and of John Conway’s On Numbers and Games,
which we are glad to say is also reappearing, the subject of combinatorial games has burgeoned
into a vast area, bringing together artificial intelligence experts, combinatorists, and computer
scientists, as well as practitioners and theoreticians of particular games such as Go, Chess,
Amazons and Konane: games much more interesting to play than the simple examples that
we needed to introduce our theory.
Just as the subject of combinatorics was slow to be accepted by many “serious” mathemati-
cians, so, even more slowly, is that of combinatorial games. But now it has achieved consid-
erable maturity and is giving rise to an extensive literature, documented by Aviezri Fraenkel
and exemplified by the book Mathematical Go: Chilling Gets the Last Point by Berlekamp
and Wolfe. Games are fun to play and it’s more fun the better you are at playing them.
The subject has become too big for us to do it justice even in the four-volume work that we
now offer. So we’ve contented ourselves with a minimum of necessary changes to the original
text (we are proud that our first formulations have so well withstood the test of time), with
additions to the Extras at the ends of the chapters, and with the insertion of many references
to guide the more serious student to further reading. And we’ve corrected some of the one
hundred and sixty-three mistakes.
We are delighted that Alice and Klaus Peters have agreed to publish this second edition.
Their great experience, and their competent and cooperative staff, notably Sarah Gillis and
Kathryn Maier, have been invaluable assets during its production. And of course we are
indebted to the rapidly growing band of people interested in the subject. If we mention one
name we should mention a hundred; browse through the Index and the References at the end of
each chapter. As a start, try Games of No Chance, the book of the workshop that we organized
a few years ago, and look out for its successor, More Games of No Chance, documenting the
workshop that took place earlier this year.
November 3, 2000
xiii
Preface
Does a book need a Preface? What more, after fifteen years of toil, do three talented authors
have to add. We can reassure the bookstore browser, “Yes, this is just the book you want!”
We can direct you, if you want to know quickly what’s in the book, to the last pages of this
preliminary material. This in turn directs you to Volume 1, Volume 2, Volume 3 and Volume 4.
We can supply the reviewer, faced with the task of ploughing through nearly a thousand
information-packed pages, with some pithy criticisms by indicating the horns of the polylemma
the book finds itself on. It is not an encyclopedia. It is encyclopedic, but there are still
too many games missing for it to claim to be complete. It is not a book on recreational
mathematics because there’s too much serious mathematics in it. On the other hand, for us, as
for our predecessors Rouse Ball, Dudeney, Martin Gardner, Kraitchik, Sam Loyd, Lucas, Tom
O’Beirne and Fred. Schuh, mathematics itself is a recreation. It is not an undergraduate text,
since the exercises are not set out in an orderly fashion, with the easy ones at the beginning.
They are there though, and with the hundred and sixty-three mistakes we’ve left in, provide
plenty of opportunity for reader participation. So don’t just stand back and admire it, work
of art though it is. It is not a graduate text, since it’s too expensive and contains far more
than any graduate student can be expected to learn. But it does carry you to the frontiers of
research in combinatorial game theory and the many unsolved problems will stimulate further
discoveries.
We thank Patrick Browne for our title. This exercised us for quite a time. One morning,
while walking to the university, John and Richard came up with “Whose game?” but realized
they couldn’t spell it (there are three tooze in English) so it became a one-line joke on line
one of the text. There isn’t room to explain all the jokes, not even the fifty-nine private ones
(each of our birthdays appears more than once in the book).
Omar started as a joke, but soon materialized as Kimberley King. Louise Guy also helped
with proof-reading, but her greater contribution was the hospitality which enabled the three
of us to work together on several occasions. Louise also did technical typing after many drafts
had been made by Karen McDermid and Betty Teare.
Our thanks for many contributions to content may be measured by the number of names
in the index. To do real justice would take too much space. Here’s an abridged list of helpers:
Richard Austin, Clive Bach, John Beasley, Aviezri Fraenkel, David Fremlin, Solomon Golomb,
Steve Grantham, Mike Guy, Dean Hickerson, Hendrik Lenstra, Richard Nowakowski, Anne
Scott, David Seal, John Selfridge, Cedric Smith and Steve Tschantz.
xiv
♥ Preface xv
No small part of the reason for the assured success of the book is owed to the well-informed
and sympathetic guidance of Len Cegielka and the willingness of the staff of Academic Press
and of Page Bros. to adapt to the idiosyncrasies of the authors, who grasped every opportunity
to modify grammar, strain semantics, pervert punctuation, alter orthography, tamper with
traditional typography and commit outrageous puns and inside jokes.
Thanks also to the Isaak Walton Killam Foundation for Richard’s Resident Fellowship
at The University of Calgary during the compilation of a critical draft, and to the National
(Science & Engineering) Research Council of Canada for a grant which enabled Elwyn and
John to visit him more frequently than our widely scattered habitats would normally allow.
And thank you, Simon!
I have heard her declare, under the rose, that Hearts was her favourite suit.
Charles Lamb, Essays of Elia, Mrs. Battle’s Opinions on Whist.
So far our compound games have been played by two players who move alternately in just
one component at a time, and the rules have ensured that they always end, the last player
to move being the winner. Now for a change of heart, let’s see what happens when we break
some of these rules.
In Chapter 9, you must move in every component, and in Chapter 10, you can move in
whatever components you like.
In Chapter 11, there are some partizan games with infinitely many positions, and some
other loopy games in which play might continue forever.
Chapter 12 deals with the rather different theory of impartial loopy games, and with
some other modifications of the impartial theory, which might allow a player to make several
consecutive moves.
Chapter 13 gives the theory of impartial games when the last player is declared to be the
loser.
xvii
-9-
If You Can’t Beat ’Em, Join ’Em!
277
278 Chapter 9. If You Can’t Beat ’Em, Join ’Em! ♥
We shall play our first few games on the 8 by 8 chessboard illustrated in Fig. 1 with a
number of horses. As his move, Left must move every horse he can two places West and one
place North or South; Right, for his part, must move every horse he can two places North and
one place East or West, as in Fig. 1. So the horses move rather like knights in Chess, but there
are several differences. Each player is limited to just 2 of the possibly 8 directions a knight
can move in; there may be arbitrarily many horses on the same square; and the same horse
is moved by both players (the horses belong to the King, not to Left or Right). Compare the
White Knight in Chapter 3.
A player will be unable to move in the game if there is any one horse which he can’t move.
According to the normal play rule he would then lose, but in this chapter we shall also give
equal treatment to the misère play rule under which he would win.
called the conjunctive compound by Smith and in ONAG, but here, for short, their join.
To move in the compound game you must make a move legal for you in every one of the
component games
G, H, K, . . .
rather than in just one, as in the sum or disjunctive compound. If you cannot do so you lose
in normal play, but win in misère play.
The maxim holds for joins of any games. When we know who starts, a game played in this
way lasts for a perfectly definite number of moves which C.A.B. Smith has called its Steinhaus
function or remoteness.
We use the term left remoteness when Left starts, and right remoteness when Right
starts. Since the turns alternate, we need only consider the Right remotenesses of Left’s options
and the left remotenesses of Right’s. You should try to leave an even remoteness (as small as
possible) for your opponent, so as to ensure that when the remoteness is reduced to zero it
will be his turn to move. This is because (in normal play)
To find the right remoteness, use the left remotenesses of the right options, GR , but still
prefer LEAST EVEN else GREATEST ODD.
280 Chapter 9. If You Can’t Beat ’Em, Join ’Em! ♥
+ + − −
RL RR (normal play) RL RR (misère play)
00 00 10 10 10 10 10 10 00 00 10 10 10 10 10 10
00 00 10 10 10 10 10 10 00 00 20 30 30 30 30 30
01 01 11 12 12 12 12 12 01 02 22 42 52 52 52 52
01 01 21 22 22 32 32 32 01 03 24 44 64 74 74 74
01 01 21 22 22 32 32 32 01 03 25 46 66 86 96 96
01 01 21 23 23 33 34 34 01 03 25 47 68 88 A8 B8
01 01 21 23 23 43 44 44 01 03 25 47 69 8A AA CA
01 01 21 23 23 43 44 44 01 03 25 47 69 8B AC BB
(a) First horse stuck loses. (b) First horse stuck wins.
+ + − −
RL RR (normal with pass) RL RR (misère with pass)
00 00 12 12 34 34 56 56 00 00 12 12 45 56 56 56
00 00 12 12 34 34 56 56 00 00 23 34 34 34 67 78
21 21 11 14 34 36 56 56 21 32 22 42 42 56 56 56
21 21 41 44 44 56 56 76 21 43 24 44 54 64 74 77
43 43 43 44 44 56 56 76 54 43 24 45 66 76 86 76
43 43 63 65 65 55 58 76 65 43 65 46 67 66 98 A8
65 65 65 65 65 85 88 86 65 76 65 47 68 89 88 B9
65 65 65 67 67 67 68 66 65 87 65 77 67 8A 9B AA
(c) First horse home wins. (d) First horse home loses.
Table 1. How Remote Are All the Horses? (A=10, B=11, C=12.)
♥ What If the First Horse to Get Stuck Wins? 281
Table 1(a) gives the left and right remotenesses (in normal play) for horses in every possible
position. See how a 0 on the left side corresponds to a position from which Left cannot move.
Figure 2 illustrates a case with larger remotenesses. Here Left’s two options have (right)
remotenesses 0 and 1; of these he prefers the only even number, 0, and adds 1 to obtain 1.
Right’s two options have (left) remoteness 1 (not much choice!) and he adds 1 to obtain 2.
Table 1(b) gives the misère play remotenesses for our game.
Since, in either case, a join of games finishes when its first component does, its remoteness
(of any kind) is the least remoteness (of the same kind) of any of the components:
282 Chapter 9. If You Can’t Beat ’Em, Join ’Em! ♥
+
RL (G ∧ H ∧ . . .) = min(R+ +
L (G), RL (H), . . .),
+ + +
RR (G ∧ H ∧ . . .) = min(RR (G), RR (H), . . .),
− − −
RL (G ∧ H ∧ . . .) = min(RL (G), RL (H), . . .),
− −
RR (G ∧ H ∧ . . .) = min(RR (G), R−
R (H), . . .).
in MISÈRE play.
The left remotenesses of the two horses are 10 and 11 (A and B) so the left remoteness
of the position as a whole is the least of these, 10. Since this is even, Left has a good move
which changes this to the ODD number 9, but the right remotenesses of these horses are 12
and 11 (C and B), minimum 11, and so from this position Right has no good move. Left’s
favorite horse is the left one, but Right’s is the right one, even though this seems further from
finishing.
♥ A Slightly Slower Join 283
really is the favorite in normal play (remoteness 4 against two 5s), but the two trailing ones
are joint favorites in misère play (remotenesses 6 as against 7).
All these games may be played on any size of board, or even on a quarter-infinite one.
Table 5 (in the Extras) gives remotenesses for this latter case.
In another version (Tables 2(c) and 2(d)) a player who cannot cut a particular piece of
cake may pass over that piece provided his opponent can still cut it. The game ends when the
first 1 by 1 cake appears, for this admits no cut by either player. So for a horizontal strip,
which cannot be cut by Rita, her remoteness is one more than Lefty’s.
In impartial Cutcakes (Tables 2(x) and 2(y)) the players must cut all the cakes, but each
may do so in either direction.
In some of these tables we have “writ large” some remotenesses which are the same for the
whole blocks of entries.
♥ Cutting Every Cake 285
Eatcakes
Remoteness tables for Eatcakes, a variant game which is more natural as a slow join (see
later), are given in the Extras (Table 7).
G, H, K, . . .
to produce a compound,
called the continued conjunctive compound in ONAG, and here, the slow join. Our previous
kind of join may be called the rapid join when we need to avoid confusion. In the slow join of
a number of games a player must move in every component he can, and the game ends only
when he cannot move anywhere.
The best tactics are a travesty of those for the rapid join — move slowly when you’re
winning, quickly when you’re losing ! The winner is anxious to savor his inexorable superiority
for as long as possible, while the loser wants to get it over with, but quick! Given who starts,
a game lasts, when played according to these cat and mouse tactics, for a perfectly definite
number of moves, called the suspense number.
We find these suspense numbers by a parody of the remoteness rules:
Another random document with
no related content on Scribd:
5. Draw out piston rod slowly, supporting it at both ends as it
leaves cylinder. Hold recoil valve in cylinder with ends of fingers.
Keep receptacle under front of cylinder to catch surplus oil.
6. Drain surplus oil from piston rod by holding vertically over
receptacle with piston down and holding buffer rod in place.
7. Rest piston rod on blocks, remove buffer-bushing locking screw,
and unscrew buffer bushing, holding piston rod by wrench on flats at
front end of rod. Have supporting blocks under both ends of rod so
that rod will not be strained.
8. Draw out buffer rod carefully.
9. Draw out recoil valve with fingers. Remaining parts can now be
easily disassembled. Buffer head is locked in place with bronze pin,
which must be driven out before head can be unscrewed.
GENERAL INSTRUCTIONS.
Filling Recoil Cylinder (Cylinder Mounted on Carriage.)
The Carriage.
Weight of carriage, complete 1950 lbs.
Weight of gun and carriage complete 2945 lbs.
Weight at end of trail carriage limbered 96 lbs.
Diameter of wheels 56 in.
Width of track 60 in.
Length of recoil 49 in.
Maximum angle of elevation 16 degrees
Maximum angle of depression 5 degrees
Maximum amount of traverse of gun on the carriage 142 mils
DESCRIPTION.
The gun is a combination of a built up and a wire wrapped gun. It
consists of a tube, a series of layers of steel wire, jacket and breech
ring. The tube extends from the rear end of the chamber to the
muzzle. Over the rear portion of the tube are wound 15 layers of
O.04 by O.25” steel wire. The jacket is fitted over the wire and the
tube, and is secured longitudinally by corresponding shoulders and
the breech ring, which is screwed over the jacket at the rear, and
secured by a set screw. The breech ring is prepared for the reception
of the breech mechanism, and is provided on the upper side with a
lug for the attachment of the hydraulic buffer.
GUN.
Weights and Dimensions.
Weight 2,688 lb.
Caliber 4.7 in.
Total length 134.92 in.
Length of bore 129.42 in.
Length of rifling 111.9 in.
Number of grooves 42
Twist Right hand.
Weight of proj. based fuse 60 lb.
Weight of proj. point fuse 45 lb.
Weight of powder charge 95 oz.
Muzzle velocity, 60 lb. proj. 1,700 ft. per second
Muzzle velocity, 45 lb. proj. 2,050 ft. per second
Maximum range 11,000 meters
THE CARRIAGE.
Weight of carriage complete 5,320 pounds
Weight of carriage and gun complete 8,068 pounds
Maximum elevation 15 degrees
Maximum depression 5 degrees
Maximum traverse 140 mils.
4.7” Rifle